Daojing Li·Hongbing Ji

Determining CO2consumption from elemental change in soil profiles developed on carbonate and silicate rocks

Daojing Li·Hongbing Ji

To further understand the roles of carbonate and silicate rocks in regulating the atmosphere/soil CO2level，the flux of CO2consumed by the chemical weathering of silicate and carbonate rocks was determined from the elemental change in soil profiles.Results showed that thechemicalweatheringofcarbonaterocksmainly occurred at the rock-regolith interface，and that the further weathering of the residua soil on the carbonate rocks was similar to that of the granite profile.Chemical weathering of the silicate rocks occurred through the whole profiles. Therefore，CO2consumed per volume by the silicate profiles［Msr（CO2）］and the residues on carbonate rocks［Mcr（CO2）］were calculated based on the elemental weathering gradients.CO2consumed by carbonate protolith［Mcp（CO2）］was calculated from the elemental change at the rock-regolith interface.The Msr（CO2）were about tens to thousands orders of magnitude greater than Mcr（CO2）.Even so，this demonstrated that the residues on carbonate rocks could be a sink of CO2on long-term scales.The Mcp（CO2）was about four times larger than Msr（CO2），which demonstrated that carbonate rocks played a more important role in regulating the CO2level than the silicate rocks did during the pedogenic process of the profiles.

Carbonate rocks·Silicate rocks·CO2consumption·Soil profiles

1　Introduction

Chemical weathering of silicate and carbonate rocks is an important source of control on the regulation of the level of atmosphericCO2，andtherebyexertscontrolonglobalclimate variations（Hartmann et al.2009；Moosdorf et al.2011；Moquet et al.2011）.To improve our understanding of that control，many studies about carbonate and silicate weathering have been carried out（Nesbitt and Young 1982；Suchet and Probst1995；Macphersonetal.2008；Lietal.2010；Heckman andRasmussen2011；Moosdorfetal.2011；Zengetal.2012）. Silicate weathering studies are conducted through relatively input—output calculations，from changes in solute compositions in pore water or ground water，or from the elemental differences between protolith and the weathered regolith（Kenoyer and Bowser 1992；White et al.1996；Murphy et al. 1998；White 2002；White etal.2005b）.Chemical weathering of silicate minerals has been proposed to be a sink of atmosphericCO2onthegeologicaltimescale（RidgwellandZeebe 2005）.Carbonatemineralsare morereactiveandsolublethan silicatemineralsandthusregulatethegeochemistryofsurface waters（Horton et al.1999；White et al.2005a）.Therefore，most studies about carbonate weathering are confined in the study of HCO3-ions in the surface water in carbonate watershed（Macpherson et al.2008；Li et al.2010；Zeng et al. 2012）.Few approaches have been taken on the soil profiles developed on carbonate rocks.Chemical weathering of carbonate minerals has been proposed to consume atmospheric CO2on a relatively short-term scale but not produce a net carbon sink on a long-term scale，because the CO2consumed bycarbonatemineralswouldgetbacktotheatmosphereduetothe formation of marine carbonate sediments（Berner et al. 1983；Berner 1991，1997；Goudie and Viles 2012）.Our previous work demonstrated that the weathering of carbonate rocks includes the weathering of carbonate minerals and the insoluble residues（Ji etal.2004a，b）.The weatheringprocess ofthe residue soils，whichis similartothe weatheringofnoncarbonate rocks，may produce a net carbon sink.This contradicted the previous belief that carbonate rocks did not produce a carbon sink on a long-term scale.This paper presents a systematic methodology for calculating CO2-consumption by the weathering of carbonate and silicate rocks basedonelementalchanges insoilprofiles，inorder tofurther understand the roles of carbonate and silicate rocks in regulating atmospheric CO2.

D.Li（?）·H.Ji

State Key Laboratory of Environmental Geochemistry，Institute of Geochemistry，Chinese Academy of Sciences，

Guiyang 550002，China

e-mail:lidaojing45@163.com

H.Ji

College of Resource Environment and Tourism，Capital Normal University，Beijing 10048，China

e-mail:jih_0000@126.com

2　Materials and methods

2.1Site description

The study of soil profiles consisted of four profiles developed on carbonate rocks and two profiles developed on silicate rocks.The carbonate profiles are located in the Guizhou Province，including those at Puchang of Suiyang County（PC），Xinpu town in Zunyi（XP），Pingba County（PB）and Hezhang County（HZB）.One silicate profile is at Hezhang County（HZX）and the other is at Gunbei village（GB）in Rongshui County，which is in the northern Guangxi Province bordering the Guizhou Province（Fig.1）.The sampling sites are all in hilly areas.The climate of the study area was during the transition zone between the East Asian and the South Asian monsoons.As the elevation increased from the east to the west，the temperature and precipitation decreased.The six profiles were selected because they were distributed from the east to the west（Fig.1）and were developed on different kinds of rocks（Table 1），based on which the effects of climate and underlying materials on weathering has been discussed.Detailed geographical and climate information of each site is tabulated in Table 1.

Fig.1 Location of study sites in Guizhou and Guangxi Province.Red pentagram indicates the locations where the soil profiles were sampled. Red dots indicate the main cities

2.2Sampling and analyzing methods

The soil profiles were obtained by digging from the bottom to the top and were classified by soil color，texture，major element content，and soil horizon characteristics（Table 1）. The profiles can be divided into three horizons:the soil horizon（A-horizon，top for farming layer），the regolith horizon（B-horizon，weathered layer）and the weathering bedrock horizon（C horizon），which can be then subdivided into the flour layer（C1-horizon），the cracked rock layer（C2-horizon）and the parent rock layer（C3-horizon）.The‘‘regolith''mentioned below refers to both the B and Ahorizons unless otherwise specified.Not all the profiles consist of all the horizons（Table 2）.Parent rock samples that were too deep to dig were collected from cores nearby. The soils were collected using a hand hammer and a small shovel，and were sampled at different intervals，depending on profile depth and soil properties.When a distinct change appeared in the soil color or texture，they were not sampled at intervals，but the soil with the different characteristics were removed.A soil sample of 5 cm in vertical width was randomly collected from each horizon.A geological map of the Guizhou Province and Guangxi Province was used to identify the parent material in the field.Each parent rock sample was identified by its texture，structure，shape and color，from at least three rock fragments.

Table 1 Geomorphological and climatic information of the soil profile

Soil samples were dried in the oven at 45°C for 2 days，with parts of each being ground to pass a sieve（mesh 200）for further analysis.The bulk density of the regolith samples was determined with the cutting-ring method，referring to the GB/T 50123-1999.The bulk density of the rock was determined with the paraffin method（GB/T 50123-1999）. The major oxides of the rock and regolith samples were analyzed by X-ray fluorescence spectrography（XRF）by using the Philips PW2404 X-ray fluorescence spectrometer and referring to the GB/T 14506.14-2010 and GB/T 14506.28-2010.The trace and rare earth element concentrations were measured using Inductively Coupled Plasma Mass Spectroscopy（HR-ICP-MS）（Element I，F(xiàn)innigan MAT Company）.The mineral composition of the carbonate rocks was analyzed using X-ray diffraction（XRD）.The mineral composition of the silicate rocks was calculated according to the CIPW norm mineral calculation method based on major oxides.

3　Results

The vertical distributions of major oxides，trace elements，chemical index of alteration（CIA）（Nesbitt and Young 1982）and the loss on ignition（LOI）are listed in Table 2.The CIA is used to estimate the weathering intensity of the regolith. An obvious increase occurs at the interface between the C horizon and B horizon for each site（Table 2）.The LOI is extremely high for parent carbonate rocks，reflecting a high carbonate content.High LOI in the regolith samples may reflect the influence of high water-bearing phases（mainly clay minerals）and organic matter（Ji et al.2004a）.

An Al2O3—CaO*+Na2O—K2O（A—CN—K）ternary diagram was drawn based on the major oxides content（where CaO*represents Ca in the silicate fraction）（Fig.2）.In the A—CN—K ternary diagram，there are two weathering trends，one parallel to the A—CN side（weathering trend 1），and the other parallel to the A—K side（weathering trend 2）（Fig.2）. Weathering trend 1 represents the process of leaching Caand Na and accumulating Al，weathering trend 2 represents the process of leaching K and accumulating Al.For the carbonate profiles（PC，HZB，XP，PB），most samples of the regolith can't be notably distinguished and almost flock together close to the area between where the illite lies and the A end.The PC，PB and HZB first experience weathering trend 1 and then trend 2.In weathering trend 1，the Ca and Na are almost completely leached from the C horizon to the B horizon.Weathering trend 2 is not very apparent，the samples of PC are close to the area where the illite lies，and those of PB and HZB tend to approach to the A end.The XP profile experiences the weathering trend 2 from the C3 horizon to the C2 horizon，and then experiences the weathering trend 1 to the joint A—K.The samples of XP lie in the area between where the illite lies and the A end.For the silicate profiles，the GB site first experiences the weathering trend 1 from the C3 horizon to the C2 horizon，and then experiences a clearly distinguishable weathering trend 2.The samples distribute uniformly close along the A—K side，and occupy the area where the illite and muscovite lies.The HZX profile only experiences weathering trend 1，and the samples distribute closely along the A—CN side，which demonstrate a wide range of weathering intensity，from unweathered to slightly weathered，to intensely weathered and finally，near the composition of pure aluminosilicate minerals.

Table 2 The vertical distributions of major oxides， trace elements， chemical index of alteration （CIA） and loss on ignition （LOI）

Table 2 continued

Table 2 continued

Fig.2 Ternary diagram of molecular proportions Al2O3—CaO*+ Na2O—K2O.Plotted are estimates of all samples of each site in this study，including unweathered bedrocks，weathered rocks，regolith and soil.Also plotted is the idealized mineral compositions.Dotted line represent the weathering trend（weatheing trend 1 and 2）.Shown at the left side is the CIA scale.Kln kaolinite，Gbs gibbsite，Ill illite，Ms muscovite，Bt biotite，Kfs K-feldspar，Pl plagioclase，Sme smectite

Table 3 Minerals of the protolith of each site

The mineral composition is purer for carbonate rocks than for silicate rocks（Table 3）.The difference is caused by different analyzing methods and the natural properties of the rock itself.The carbonate rocks are composed ofnearly pure calcite or dolomite.The mineral composition of silicate rocks is complex.For HZX（basalt），its main minerals include feldspar，pyroxene and iron oxides.The parent rock at the GB site（granite）is composed mainly by quartz，albite and orthoclase（Table 3）.

Fig.3 Elemental mobilities at each profiles.τjis the mass transfer coefficient from Eq.（1）

4　Discussion

4.1Weathering process of carbonate and silicate rocks

As shown in the A—CN—K diagram（Fig.2），the weathering process of carbonate profiles（PC，HZB，XP，PB）can be dividedintotwostages.Theweatheringreactionmainlyhappens at the first stage in the C horizon（including C1，C2 and C3），with the almost completely leaching of Ca and Na and the accumulationofAl.Thesecondstageisthefurtherweathering of insoluble residua，which is similar to the weathering of the GB profile，with the leaching ofK and accumulation of Al.At the silicate profiles（HZX and GB），weathering happens more gradually throughout the whole profile，than at the carbonate profiles.The difference comes from the different protolith compositions.Themainmineralsincarbonaterocksarecalcite or dolomite（Table 3）.The high solubility of calcite and dolomiteallowthemtobeeasilyleachedoutattheinitialstage，especially in a climate with high precipitation（Table 1）.The mainmineralsinsilicaterocksinthisstudyarequartz，feldspar and pyroxene；their weathering resistances are stronger than calcite and dolomite's，so，weathering reactions take place more gradually along the whole profiles.

4.2Elemental mobility

The weathering intensity can be described by comparing the element concentrations of the regolith to those of the original protolith.This ratio is affected by the gainsand losses of other components and the compaction or dilation of the regolith.Solving such a problem requires comparing the ratio of the mobile component，j，to the ratio of an additional inert component，i.

Fig.4 Relationships between Al2O3and TiO2，Al2O3and Fe2O3in soil profiles

τjis the mass transfer coefficient（Brimhall and Dietrich，1987）.The value of τj=0 denotes no mobility，τj=-1 indicates complete leaching out and τj＞0 denotes external additions.Refractory elements such as Zr，Ti and Nb are commonly used as conservative components，Ci，in Eq.（1）（Brimhall et al.1991；Merritts et al.1991；White et al. 1998；Riebe et al.2001）.In this study，we use Ti as the inert component.The original protolith component is that of the sample in the C horizon for each site.

For carbonate rocks（PC，HZB，XP，PB），the main minerals are calcite and dolomite（Table 3）；the weathering intensity of calcite is reflected in the mobility of Ca and the weathering intensity of dolomite is reflected in mobility of Ca and Mg.For silicate rocks，the minerals are more complex:the weathering intensity of plagioclase is reflected in the mobility of Na，Ca and Sr and the weathering intensity of K-feldspar is reflected in the mobility of K，Rb and Ba.Al，F(xiàn)e and Mg concentrations reflect the weathering of smectite and the formation of secondary kaolinite，gibbsite and Fe oxides（White et al. 2008）.

As shown in Fig.3，in the profiles developed on carbonate rocks（PC，HZB，XP，PB），the Mg，Ca，Na and Sr are totally lostintheBandAhorizons.TheSrcanreplaceCaand，when contained in calcite，the similar τjdepth distribution of Mg，Ca and Sr demonstrates carbonate and dolomite weathering. The Na mobility has a similar depth change as that ofMg，Ca and Sr because of its mobile character in carbonate weathering.The higher τjof Na that appears at the C2 horizon of the PC profile demonstrates the slower weathering rate of the albite composed in the bedrock（Table 3）.The higher τjof Mg，Ca，Na and Sr in the C horizon of the XP and PB than that of the PC and HZB is caused by the slower dissolution rate of dolomite compared to the dissolution rate of calcite.

At the profiles developed on the silicate rocks，the weathering process of basalt（HZX）isdifferentfrom that of granite（GB）.AtHZX，theelementmobilitybarelychangesattheC2 horizon，with the largest variation happening at the C2-B（rock-regolith）interface.The Mg，Ca，Na and Sr are greatly depletedattheinterfaceandarethenalmostunchangedinthe BandAhorizon，reflectingtheweatheringofthefeldsparand pyroxene that mainly occurred at the C2-B interface.At the GB site，Ca and Na are almost completely lost，reflecting the intense weatheringofalbiteandanorthite.TheAl，F(xiàn)e and Mg mobility change synchronously，reflecting the weathering of muscovite and illite.The gradually decreasing τjof K and Ba reflects the weatheringofK-feldspar.The decreased mobility of K，Ba and Rb relative to Na and Ca reflect the slower weathering of K-feldspar relative to plagioclase（Nesbitt and Young，1984；White et al.2001）.

At both the carbonate and silicate profiles，the Al and Fe mobility change synchronously，reflecting their similar geochemistry behavior in the study profiles.Except for the GB profile，F(xiàn)e and Al accumulation is accompanied by the loss of mobile elements.The fact that Fe is not accumulated relative to protolith in the GB profile may result from the lack of clay in the environment，which is in favor of the formation of iron oxides in comparison to other profiles.The less clay in the GB is reflected in the lower LOI（Table 2）. The extremely high amount of precipitation at the GB site（Table 1）may also contribute to the leaching of the Fe.

4.3Determining the sink of CO2based on elemental weathering gradients in regolith

In studies on CO2consumption by chemical weathering，one method is to use the solid-state elemental changes between protolith and regolith.However，the allochthonous components carried by wind or water present a limit for the application of the mass balance approach when calculating CO2consumption（Schellmann 1989；Maynard 1992）.It's importanttodeterminewhethertheallochthonouscomponentshave affectedthecomponentintheprofile.TheelementsAl，Tiand Fe are the less mobile elements，and there is a positive correlation between Al2O3and TiO2，as well as Al2O3and Fe2O3（Fig.4），reflecting the characteristics of the in situ weathering（Young and Nesbitt，1998）.The synchronous τjchanges of the element groups（Al—Fe，Al—Fe—Mg and Mg—Ca—Na——Sr）discussed in Sect.4.2 also demonstrate the insituweatheringofthestudyprofiles.Theinsituweathering of the carbonate profiles in the Guizhou province is also demonstratedindetailinother papers（Jietal.2004a，b；Feng etal.2009）.So，mass changes can becalculateddirectlyfrom the differences between elemental compositions in the initial protolith and the weathered regolith.

In the simplest scenario for steady-state weathering，a mobile element，which is not incorporated into secondary precipitates，linearly decreases with decreasing depth from an initial concentration C0at depth z1to a concentration Cwat a shallower depth z0（Fig.5）（White，2002）.The elemental concentrations are commonly reported as mol kg-1. In order to make direct comparisons of element mobility due to weathering，element change Δm（mol m-3）is defined in terms of a unit volume（White，2002）such that

Fig.5 Schematic showing the distributions of a mobile element in a weathering regolith.Cwdefines the weathered concentration at shallow depth z0，and C0is the initial protolith concentration at depth z1.The weathering gradients bsdescribe the slope of the linear gradients

where ρwis the regolith bulk density（g cm-3），C0（mol kg-1）is the initial elemental concentration shown in Fig.5，which is assumed to be that of the protolith，and Cw（mol kg-1）is the elemental concentration at the weathered regolith.Under closed system conditions，the difference in concentrations，C0—Cw，reflects the total mass change occurring over the entire time of pedogenesis.The concentration Cwcan't be determined directly from themeasured elemental concentration C，because C is also affected by the gains and losses of the other components and the compaction or dilation of the regolith.The element Ti is chosen as the inert component again.The normalized weathering concentration Cw，is obtained from the measured concentration C by the relationship

where I0（mol kg-1）is the concentration of the inert element in the protolith and Iwis the concentration in the weathered regolith（White，2002）.The slope of the gradient in Fig.5 is defined as bs（m kg mol-1）.Therefore，Cw-C0=Δz/bs.Substituting this relationship into Eq.（2）results in the expression

As discussed in Sect.4.2，the Mg，Ca，Na，K is the main mobile element in most minerals of the carbonate and silicate rocks，and their concentration change will reflect the chemical weathering intensity，thus reflecting the drawdown of atmospheric CO2.In this study，we ignore the role of organic acid and sulfuric acid that may take part in the chemical weathering，and regard the CO2as the only acid source that participates in the weathering reaction.

The weathering reaction that happened at the carbonate profiles at the initial stage weremainlythe leachingof calcite and dolomite demonstrated in the following equations:

Thefurtherweatheringreactionsoftheresiduesoncarbonate rocks are mainly the reactions of non-carbonate minerals.

The main chemical weathering reactions that happened at the silicate profiles based on their main minerals are mainly the weathering of albite，anorthite，orthoclase and pyroxene as follows

（Sak et al.2004）

From Eq.（5）to（10），it can be seen that no matter what chemical reaction happened，for carbonate minerals，two mole of base cation charge consume 1 mol of CO2.For non-carbonate minerals，one mole of base cation charge consumes 1 mol of CO2.The CO2captured per volume M（CO2）（mol cm-1）by carbonate and non-carbonate mineral weathering is defined as

Fig.6 Normalized concentrations of Ca（Cw）plotted as a function of depth in regolith.The lines are the linear regression fits describing the weathering gradients bs

There is a certain deviation in the calculation of M（CO2）based on the elemental change，because CO2isn't the only acid source that participates in the weathering reaction. Also，not all elements that reacted with CO2in the profiles are included.Nevertheless，the results can be used tocompare the different capabilities of capturing CO2between the carbonate and silicate profiles in the study area.

Fig.7 Normalized concentrations of Na（Cw）plotted as a function of depth in regolith.The lines are the linear regression fits describing the weathering gradients bs

Fig.8 Normalized concentrations of K（Cw）plotted as a function of depth in regolith.The lines are the linear regression fits describing the weathering gradients bs

Fig.9 Normalized concentrations of Mg（Cw）plotted as a function of depth in regolith.The lines are the linear regression fits describing the weathering gradients bs

4.3.1Calculation of CO2captured per volume of regolith Mregolith（CO2）

The normalized weathering concentration，Cw，of Ca，Na，K，and Mg are plotted as functions of depth in Figs.6，7，8， and 9 respectively.The points plotted for PC，XP，PB and HZB are confined in the B and A horizons.The HZX and GB are confined in C2，B and A horizons.The linear regression equations and the correlation coefficient（R2）drawn from Figs.6，7，8，and 9 are tabulated in Table 4.At the PB site，the Ca increases slightly in the shallower soils（Fig.6），which results in negative gradients（Table 4）. Such an increase may be attributed to the upward biologicpumping via plant roots and the subsequent recycling in the shallow soils（White et al.2009；White 2014）.Because the Cwof Ca in the PB regolith is relatively small in comparison to those in the other sites（Fig.6），it is acceptable to not use the Ca concentration in the calculation of M（CO2）. The Na didn't show any linear fit to the data at the PC site（Table 4）.To clarify the reason for this，the surface soil was analyzed by XRD，and the semiquantitative calculation of the mineral composition shows that the surface soil of the PC was composed of albite（2.82%），hornblende（1.16%），quartz（88.35%），smectite（1.85%），illite（2.83%），kaolinite（1.87%）and iron oxides（1.12%）.Albite and hornblende are primary minerals.Albite appeared in the bedrock of the PC（Table 4）.Even so，the presence of it at the surface is likely from the allochthonous materials，as albite is easily weathered at deep depths.Hornblende is not present in the bedrock，demonstrating the deposition of the allochthonous materials.Except for albite and hornblende，the others are secondary minerals，which are usually found in the weathering residues of carbonate rocks（Ji et al.2004a，b）.Primary minerals that are not present in the bedrock also appeared at the surface of the PB［albite（0.52%）and hornblende（1.30%）］and HZB sites［albite（1.31%）］.This may explain the lower linear correlation coefficient for Na at the PC，PB and HZB sites. The Cwof Na in carbonate regolith is small（about 0—0.002 mol kg-1）（Fig.7），and it is about ten orders of magnitudes smaller than that of Ca，K and Mg.Therefore，the effect of the allochthonous materials is neglected and the Na concentrations of the PC，PB and HZB were not included in the calculation of M（CO2）.The bsin Eq.（4）is the slope of the linear regression equations and the ρwis the average value of the bulk density of each horizon（Table 5）.The Mregolith（CO2）was calculated according to Eq.（12）.

Table 4 The linear regression equations and the correlation coefficient （R2） drew from Figs. 6， 7， 8 and 9

The results show that the Mregolith（CO2）of the silicate regolith Msr（CO2）is tens to thousands orders of magnitude greater than that of the carbonate regolith Mcr（CO2）（Table 5）.This occurs because the elements are greatly leached in the C horizon of the carbonate profiles and the elements that remain in the carbonate regolith are significantly less than that in the silicate regolith.Though the amount of CO2consumed by the insoluble residuals on carbonate rocks is small，it can't be denied that the chemical weathering of them may produce a net sink of CO2in the long-term scale.

For the carbonate profiles，the Mcr（CO2）is in the order of PC?XP＞PB≈HZB.The average CIA value of the carbonate regolith is PC（86），XP（89），PB（94），HZB（93），demonstratingaweatheringintensityorderof PB≈HZB＞XP＞PC，which is in the reverse when compared with the order of Mcr（CO2）.This is also related to the two-stage weathering process of the carbonate rocks；higher weathering intensity resulted in less elements left in the carbonate regolith，so the weathering of the regolith would consume less CO2.The PC profile belongs to the cambisols in soil classification and weaker development of the profile lead more elements in the regolith.Therefore，the Mcr（CO2）ofPCismuchlargerthantheotherthreecarbonate profiles.For the silicate profiles，the HZX（CIA=92）consumes more CO2than the GB（CIA=84），demonstrating that the profile with the higher weathering intensity consumes more CO2.This is consistent with the weathering process of silicate.

Table 5 The elemental change（Δm）of Ca，Na，K and Mg and the CO2captured per volume of regolith Mregolith（CO2）

Table 6 The elemental change（Δm）of Ca and Mg in carbonate rocks and the CO2captured per volume of carbonate protolith Mcp（CO2）

Usually，warmer and wetter climates yield higher weathering intensity（White 2002；White and Blum 1995），thus possibly leading to a higher carbon sink.The studied profiles were distributed from the west to the east（Fig.1），with relatively warmer and wetter climates in the east.From Table 1，it can be seen that the mean annual temperature is GB＞PB＞PC＞XP＞HZB=HZXandthemeanannual precipitationisGB＞PB＞PC=XP＞HZB=HZX，based on the Mregolith（CO2）of each site.Itseems that there is no correlation between the climate and the Mregolith（CO2）. However，we cannot conclude that the climate doesn't show any impact on the carbon sink in the study area，because the climatic signals are obscured by the lithologic difference and the pedogenic time.

4.3.2Calculation of CO2captured per volume of carbonate protolith Mcp（CO2）

The elemental change between the protolith and the regolith in the silicate profiles is small and the weathering penetrates from the soil horizon to the protolith gradually，so the elemental change calculated based on weathering gradients can reflect the change of the whole silicate profile.However，as discussed in Sects.4.1 and 4.2，the weathering of the carbonate protolith happened mainly at the rock-regolith interface，with the almost complete loss of Ca and Mg by the reactions of Eq.（5）and Eq.（6）.To determine the element change at the interface，the C0，Cw，and ρwin Eq.（4）are regarded as the elemental concentration of the protolith，and the elemental concentration of the bottom layer of the regolith and the bulk density of the protolith respectively.The CO2captured per volume of carbonate rock Mcp（CO2）is calculated according to Eq.（11）.Based on this principle，the calculated Mcp（CO2）are about four times larger than Msr（CO2）（Table 6）.The results demonstrate that during the entire time of soil development，the carbonate rocks play a more important role than the silicate rocks in regulating the level of CO2.

5　Conclusions

This work compared the weathering process of carbonate and silicate rocks and determined the CO2captured during this process.Due to the extremely high dissolution rates of calcite and dolomite，the weathering of the carbonate rocks mainly occurred at the rock-regolith interface，with leaching of most mobile element.The further weathering of the insoluble residues on the carbonate rocks is similar to that of granite profile（GB）.The silicate weathering，especially the weathering of granite，proceeded more gradually than the carbonate weathering.Therefore，the mass changeoccurred at the regolith of the carbonate profiles and the silicate profiles are calculated based on the weathering gradients.The mass change that occurred at the rock-regolith interface of the carbonate profiles was calculated by the elemental differences between the protolith and the bottom regolith layer.The CO2consumed by silicate profiles are tens to thousands orders of magnitude greater than that by carbonate regolith.Even so，it demonstrates that the carbonate regolith can be a sink of CO2on the long-term scale.The CO2consumed by carbonate rocks are about four times larger than that by silicate profiles，implying that carbonate rocks play a more important role than silicate rocks in regulating the level of CO2during the pedogenic process of the profiles.

AcknowledgmentsThis work was jointly supported by the National Natural Science Foundation of China（NSFC）grants（No.41073096 and 40473051），National Key Basic Research Program of China（2013CB956702）and the Hundred Talents Program of the Chinese Academy of Sciences.We thank Gong G.H.for the XRD analysis and Hu J.for assistance in the HR-ICP-MS determination.

Berner RA（1991）A model for atmospheric CO2over Phanerozoic time.Ar J Sci 291:339—376

Berner RA（1997）Weathering，plants and long-term carbon cycle. Geochim Cosmochim Acta 56:3225—3231

Berner RA，Lasaga AC，Garrels RM（1983）The carbonate—silicate geochemical cycle and its effect on atmospheric CO2.Am J Sci 283:641—683

Brimhall GH，Dietrich WE（1987）Constitutive mass balance relations between chemical-composition，volume，density，porosity，and strain in metasomatic hydrochemical systems-results on weathering and pedogenesis.Geochim Cosmochim Acta 51:567—587

Brimhall GH，Lewis CJ，F(xiàn)ord C，Bratt J，Taylor G，Warin O（1991）Quantitaive gechemical approach to pedogenesis-importance of parent material reduction，volumetric expansion，and elian influx in lateritization.Geoderma 51:51—91

Fedo CM，Nesbitt HW，Young GM（1995）Unraveling the effects of potassium metasomatism in sedimentary rocks and paleosols，with implications for paleoweathering conditions and provenance.Geology 23:921—924

Feng JL，Cui ZJ，Zhu LP（2009）Origin of terra rossa over dolomite on the Yunnan-Guizhou Plateau，China.Geochem J 43:151—166

Garrels RM，Mackenzie FT（1971）Evolution of sedimentary rocks. Norton and Co.Inc.，New York，p 397

GB/T 14506.14-2010.Silicate rock chemical analytical procedure part 14:determination of ferrous oxide（in Chinese）

GB/T 14506.28-2010 Silicate rock chemical analytical procedure part 28:determination of 16 primary and secondary components（in Chinese）

GB/T 50123-1999.Standard for soil test methods part 5:determination of soil density（in Chinese）

Goudie AS，Viles HA（2012）Weathering and the global carbon cycle: Geomorphological perspectives.Earth Sci Rev 113:59—71

Hartmann J，Jansen N，Du¨rr HH，Kempe S，Ko¨hler P（2009）Global CO2-consumption by chemical weathering:What is the contribution of highly active weathering regions？Global Planet Change 69:185—194

Heckman K，Rasmussen C（2011）Lithologic controls on regolith weathering and mass flux in forested ecosystems of the southwestern USA.Geoderma 164:99—111

Horton TW，Chamberlain CP，F(xiàn)antle M，Blum JD（1999）Chemical weathering and lithological controls of water chemistry in a high-elevation river systems:Clark's Fork of the Yellowstone River，Wyoming and Montana.Water Resour Res 35:1643—1655

Huh YS（2003）Chemical weathering and climate—a global experiment:a review.Geosci J 7:277—288

Ji HB，Wang SJ，Ouyang ZY，Zhang S，Sun CX，Liu XM，Zhou DQ（2004a）Geochemistry of red residua underlying dolomites in karst terrains of Yunnan-Guizhou Plateau I.The formation of the Pingba profile.Chem Geol 203:1—27

Ji HB，Wang SJ，Ouyang ZY，Zhang S，Sun CX，Liu XM，Zhou DQ（2004b）Geochemistry of red residua underlying dolomites in karst terrains of Yunnan-Guizhou Plateau II.The mobility of rare earthelementsduringweathering.ChemicalGeology. 203:29—50

Kenoyer GJ，Bowser CJ（1992）Groundwater chemical evolution in a sandy silicate aquifer in northern Wisconsin，2.Reaction modeling.Water Resour Res 28:591—600

Li SL，Liu CQ，Li J，Lang YC，Ding H，Li LB（2010）Geochemistry of dissolved inorganic carbon and carbonate weathering in a small typical karstic catchment of Southwest China:isotopic and chemical constraints.Chem Geol 277:301—309

Macpherson GL，Roberts JA，Blair JM，Townsend MA，F(xiàn)owle DA，Beisner KR（2008）Increasing shallow groundwater CO2and limestone weathering，Konza Prairie，USA.Geochim Cosmochim Acta 72:5581—5599

Maynard JB（1992）Chemistry of modern soils as a guide to interpreting Precambrian paleosoils.J Geol 100:279—289

McLennan SM （1993）Weathering and global denudation.J Geol 101:295—303

Merritts DJ，Chadwick OA，Hendricks DM （1991）Rates and processes of soil evolution on uplifted marine terraces，northern California.Geoderma 51:241—275

Moosdorf N，Hartmann J，Lauerwald R，Hagedorn B，Kempe S（2011）Atmospheric CO2consumption by chemical weathering in North America.Geochim Cosmochim Acta 75:7829—7854

Moquet J-S，Crave A，Viers J，Seyler P，Armijos E，Bourrel L，Chavarri E，Lagane C，Laraque A，Casimiro WSL，Pombosa R，Noriega L，Vera A，Guyot J-L（2011）Chemical weathering and atmospheric/soil CO2uptake in the Andean and Foreland Amazon basins.Chem Geol 287:1—26

Murphy SF，Brantley SL，Blum AE，White AF，Dong HL（1998）Chemical weathering in a tropical watershed，Luquillo mountains，Puerto Rico:II.Rate and mechanism of biotite weathering. Geochim Cosmochim Acta 62:227—243

Nesbitt HW，Young GM（1982）Early proterozoic climates and plate motions inferred from major element chemistry of lutites.Nature 299:715—717

Nesbitt HW，Young GM（1984）Prediction of some weathering trends of plutonic and volcanic-rocks based on thermodynamic and kineticconsiderations.GeochimCosmochimActa 48:1523—1534

Ridgwell A，Zeebe RE（2005）The role of the global carbonate cycle in the regulation and evolution of the Earth system.Earth Planet Sci Lett 234:299—315

Riebe CS，Kirchner JW，Granger DE，F(xiàn)inkel RC（2001）Strong tectonic and weak climatic control of long-term chemical weathering rates.Geology 29:511—514

Sak PB，F(xiàn)isher DM，Gardner TW，Murphy K，Brantley SL（2004）Rates of weathering rind formation on Costa Rican basalt. Geochim Cosmochim Acta 68:1453—1472

Schellmann W（1989）Allochthonous surface alteration of Nilaterites.Chem Geol 74:351—364

Suchet PA，Probst J-L（1995）A global model for present-day atmospheric/soil CO2consumption by chemical erosion of continental rocks（GEM-CO2）.Tellus B 47:273—280

White AF（2002）Determining mineral weathering rates based on solid and solute weathering gradients and velocities:application to biotite weathering in saprolites.Chem Geol 190:69—89

White AF（2014）Chemical weathering of pleistocene glacial outwash sediments:a comparison of contemporary and Long-term rates for soils and groundwaters.Aquat Geochem 20:141—165

White AF，Blum AE（1995）Effects of climate on chemical weathering in watersheds.Geochim Cosmochim Acta 59（9）:1729—1747

White AF，Blum AE，Schulz MS，Bullen TD，Harden JW，Peterson ML（1996）Chemical weathering rates of a soil chronosequence on granitic alluvium:1.Quantification of mineralogical and surface area changes and calculation of primary silicate reaction rates.Geochim Cosmochim Acta 60:2533—2550

White AF，Blum AE，Schulz MS，Vivit DV，Stonestrom DA，Larsen M，Murphy SF，Eberl D（1998）Chemical weathering in a tropical watershed，Luquillo mountains，Puerto Rico:I.Longterm versus short-term weathering fluxes.Geochim Cosmochim Acta 62:209—226

White AF，Bullen TD，Schulz MS，Blum AE，Huntington TG，Peters NE（2001）Differential rates of feldspar weathering in granitic regoliths.Geochim Cosmochim Acta 65:847—869

White AF，Schulz MS，Lowenstern JB，Vivit DV，Bullen TD（2005a）The ubiquitous nature of accessory calcite in granitoid rocks: implications for weathering，solute evolution，and petrogenesis. Geochim Cosmochim Acta 69:1455—1471

White AF，Schulz MS，Vivit DV，Blum AE，Stonestrom DA，Harden JW（2005b）Chemical weathering rates of a soil chronosequence on granitic alluvium:III.Hydrochemical evolution and contemporary solute fluxes and rates.Geochim Cosmochim Acta 69:1975—1996

White AF，Schulz MS，Vivit DV，Blum AE，Stonestrom DA，Anderson SP（2008）Chemical weathering of a marine terrace chronosequence，Santa Cruz，California I:interpreting rates and controls based on soil concentration-depth profiles.Geochim Cosmochim Acta 72:36—68

White AF，Schulz MS，Stonestrom DA，Vivit DV，F(xiàn)itzpatrick J，Bullen TD，Maher K，Blum AE（2009）Chemical weathering of a marine terrace chronosequence，Santa Cruz，California.Part II: solute profiles，gradients and the comparisons of contemporary and long-term weathering rates.Geochim Cosmochim Acta 73:2769—2803

Young GM，Nesbitt HW（1998）Processes controlling the distribution of Ti and Al in weathering profiles，siliciclastic sediments and sedimentary rocks.J Sediment Res 68:448—455

Zeng C，Gremaud V，Zeng HT，Liu ZH，Goldscheider N（2012）Temperature-driven meltwater production and hydrochemical variations at a glaciated alpine karst aquifer:implication for the atmospheric CO2sink under global warming.Environ Earth Sci 65:2285—2297

10 September 2014/Revised:4 November 2014/Accepted:11 November 2014/Published online:7 February 2015 ?Science Press，Institute of Geochemistry，CAS and Springer-Verlag Berlin Heidelberg 2015